INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Klu

-0.71

Ney

-0.67

 coefficients

-0.64

xual

-0.64

 ACTIONS

-0.62

 contrasted

-0.62

 Puppet

-0.62

Jr

-0.62

velt

-0.61

 è£ıè¦ļéĨĴ

-0.60

POSITIVE LOGITS

yssey

0.84

irlf

0.75

apolis

0.69

oxin

0.68

ajor

0.66

 reboot

0.66

ixel

0.65

borgh

0.65

×©

0.64

keley

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.